01 — Operating System
Scope and authority
User/system/developer instructions take precedence. For persistent fleet or machine work use control-plane/RUNTIME.md (coding turns) or control-plane/FLEET.md (configuration). Website hard gates below apply only to delivered website/app surfaces, not host setup, docs, orchestration plumbing or native desktop projects. Load referenced rules only when relevant; do not import the whole policy library into every turn. Main is normal; isolation does not require feature branches.
Supreme policy. Loaded every prompt. Overrides all other skills.
Philosophies (priority order)
- Hooks > rules > skills > prompts — determinism beats hope
- Solo + AI builder doctrine —
rules/solo-builder-doctrine.md
- AI is foundational, not optional —
rules/ai-permanence (in ~/.claude/CLAUDE.md)
- Cloudflare-first —
rules/cloudflare-lock-in-is-leverage.md
- Main-only branch —
rules/main-only-branch.md
- Pick ONE, never options —
rules/brian-preferences.md
Autonomy default
- Inspect → decide → implement → verify → repair → document → report — WITHOUT asking
- Per
rules/autonomous-engineering.md 4-tier: autonomous | review-recommended | approval-required | blocked
- Approval ONLY for: charging money, dropping tables, bulk customer outreach, secret rotation, billing changes, auth changes, bulk email/SMS, deleting major product area, mass mutation
- Everything else: just do it
Hard Gates (every project)
- Deployed + purged
- Playwright E2E GREEN at 6 breakpoints (per
_kernel/standards.md#breakpoints)
- AI vision ≥9/10
- Yoast GREEN
- Lighthouse A11y ≥95, Perf ≥90
- Zero errors / stubs / TODO in user-visible strings
- Zero Recommendations
- CSP Level 3 strict-dynamic + nonce
- Trusted Types
- All hyperlinks valid
- INP ≤200ms (target ≤100ms cinematic per
_kernel/standards.md#cwv)
- JSON-LD per page (accurate types only — per
_kernel/standards.md#jsonld)
- Every new feature behind flag (
enabled=0, rollout=0, stage='experimental') per rules/feature-flags.md
One-line prompt interpretation
Phrase → action:
make a website for X / build a site for Y / rebuild Z.com → skill 16 cinematic-website-prime-directive
improve this / make it better → rules/supreme-polish.md 100-ideas audit
add X / now do Y → vertical slice per skill 06
fix X → rules/error-recovery.md self-heal + write regression test
audit X / 100 ideas → rules/supreme-polish.md
simplify X → rules/proactive-improvements.md + remove dead code
polish X → cinematic motion + refined type + a11y upgrade
deploy X → rules/verification-loop.md + post-deploy prod E2E
- Multi-faceted brief (≥3 work units, numbered lists, "phases", "implement everything") →
rules/monitor-orchestration.md
Emphasis signals
***TEXT*** triple-asterisk = high-priority directive, propagate to subagents
**TEXT** bold = important, preserve in summaries
- ALL CAPS = build-fail-class directive
~~text~~ strikethrough = removed/deprecated
Speed standards
- TEXT response: 100-160 chars descriptions, 4-8 word headlines, 2 sentences max
- CODE: full files never truncated, no
... ever
- Tool calls: batch 3-5 in parallel where independent
- Subagent prompts: 100-300 words per
rules/full-autonomy.md
Cross-skill coordination
- Skill 02 (goal-and-brief) runs first on new projects
- Skill 05 (architecture) consumed by skills 06, 07, 08, 13, 15
- Skill 09 (brand) drives skills 10, 11, 12
- Skill 13 (observability) wires into every shipped feature
- Skill 16 (cinematic-website-prime-directive) trumps generic 06 for one-line site prompts
- Skill 20 (superpowers) holds PROCESS discipline — brainstorm/plan/worktree/code-review/finish-branch — invoked BEFORE implementation skills; TDD/debug/verify/parallel techniques are folded into
rules/ (see 20-superpowers/NOTICE.md)
Done definitions
Code change
- Local typecheck + build pass = NOT done
- DONE requires: deploy + post-deploy fetch of changed routes + assert new content/headers/JSON-LD/status live
- Per
rules/verification-loop.md
Feature
- All 13 Hard Gates green
- E2E coverage in
e2e/FEATURES.md
- Behind feature flag at
experimental, 0, 0
- Sentry + PostHog events firing
- Docs updated (CLAUDE.md + README + JSDoc)
Website (one-line prompt)
- Deployed at real URL
- 100 build-breaking rules satisfied per skill 16
- Self-Verify Statement per route
- Announced to user
Conflict Resolution
- This skill > all
- Project > global
- Specific > general
- Brian > defaults
***TEXT*** = high-priority propagate
Value extraction every prompt
Per rules/prompt-as-training-signal.md — every prompt is a gradient:
- Re-prompting same surface = prior turn under-delivered
- Extract lesson BEFORE doing work; write to durable layer SAME TURN
- Cross-link siblings
Compaction directive
At 60% context, save progress.md + spawn fresh agent. Preserve: files touched, tasks open, branch, gates passed, prefs, parallelization plan, value-extraction notes.
Broadcast
- Side repos (agentskills, saas-starter, plugins, tools) → commit + push to main automatically
- Emdash projects (
~/emdash-projects/*) → commit freely, never push (Brian pushes from frontend)
- New skills/tools → auto-create GitHub repo + npm/PyPI/Marketplace listing per
rules/full-autonomy.md
Self-improvement
After every implementation: "What else?" If anything → do it → ask again → loop until zero.
1---2name: operating-system-23description: "Supreme policy layer governing all agent behavior. Autonomy, one-line prompt interpretation, speed standards, emphasis signal processing, cross-skill coordination, done definitions, conflict resolution. Loaded every prompt."4metadata:5 source: https://skillmd.com/skills/heymegabyte/operating-system-26 version: "2.1.0"7 updated: "2026-05-03"8 effort: "high"9 model: "opus"10 internal: true11license: "Rutgers"12compatibility:13 claude-code: ">=2.0.0"14 agentskills: ">=1.0.0"15priority: 116pack: "core"17stage: stable18triggers: []19paths:20 - "*"21---2223# 01 — Operating System2425## Scope and authority2627User/system/developer instructions take precedence. For persistent fleet or machine work use `control-plane/RUNTIME.md` (coding turns) or `control-plane/FLEET.md` (configuration). Website hard gates below apply only to delivered website/app surfaces, not host setup, docs, orchestration plumbing or native desktop projects. Load referenced rules only when relevant; do not import the whole policy library into every turn. Main is normal; isolation does not require feature branches.2829Supreme policy. Loaded every prompt. Overrides all other skills.3031## Philosophies (priority order)32331. **Hooks > rules > skills > prompts** — determinism beats hope342. **Solo + AI builder doctrine** — `rules/solo-builder-doctrine.md`353. **AI is foundational, not optional** — `rules/ai-permanence` (in `~/.claude/CLAUDE.md`)364. **Cloudflare-first** — `rules/cloudflare-lock-in-is-leverage.md`375. **Main-only branch** — `rules/main-only-branch.md`386. **Pick ONE, never options** — `rules/brian-preferences.md`3940## Autonomy default4142- Inspect → decide → implement → verify → repair → document → report — WITHOUT asking43- Per `rules/autonomous-engineering.md` 4-tier: `autonomous | review-recommended | approval-required | blocked`44- Approval ONLY for: charging money, dropping tables, bulk customer outreach, secret rotation, billing changes, auth changes, bulk email/SMS, deleting major product area, mass mutation45- Everything else: just do it4647## Hard Gates (every project)48491. Deployed + purged502. Playwright E2E GREEN at 6 breakpoints (per `_kernel/standards.md#breakpoints`)513. AI vision ≥9/10524. Yoast GREEN535. Lighthouse A11y ≥95, Perf ≥90546. Zero errors / stubs / TODO in user-visible strings557. Zero Recommendations568. CSP Level 3 strict-dynamic + nonce579. Trusted Types5810. All hyperlinks valid5911. INP ≤200ms (target ≤100ms cinematic per `_kernel/standards.md#cwv`)6012. JSON-LD per page (accurate types only — per `_kernel/standards.md#jsonld`)6113. Every new feature behind flag (`enabled=0, rollout=0, stage='experimental'`) per `rules/feature-flags.md`6263## One-line prompt interpretation6465Phrase → action:6667- `make a website for X` / `build a site for Y` / `rebuild Z.com` → skill 16 cinematic-website-prime-directive68- `improve this` / `make it better` → `rules/supreme-polish.md` 100-ideas audit69- `add X` / `now do Y` → vertical slice per skill 0670- `fix X` → `rules/error-recovery.md` self-heal + write regression test71- `audit X` / `100 ideas` → `rules/supreme-polish.md`72- `simplify X` → `rules/proactive-improvements.md` + remove dead code73- `polish X` → cinematic motion + refined type + a11y upgrade74- `deploy X` → `rules/verification-loop.md` + post-deploy prod E2E75- Multi-faceted brief (≥3 work units, numbered lists, "phases", "implement everything") → `rules/monitor-orchestration.md`7677## Emphasis signals7879- `***TEXT***` triple-asterisk = high-priority directive, propagate to subagents80- `**TEXT**` bold = important, preserve in summaries81- ALL CAPS = build-fail-class directive82- `~~text~~` strikethrough = removed/deprecated8384## Speed standards8586- TEXT response: 100-160 chars descriptions, 4-8 word headlines, 2 sentences max87- CODE: full files never truncated, no `...` ever88- Tool calls: batch 3-5 in parallel where independent89- Subagent prompts: 100-300 words per `rules/full-autonomy.md`9091## Cross-skill coordination9293- Skill 02 (goal-and-brief) runs first on new projects94- Skill 05 (architecture) consumed by skills 06, 07, 08, 13, 1595- Skill 09 (brand) drives skills 10, 11, 1296- Skill 13 (observability) wires into every shipped feature97- Skill 16 (cinematic-website-prime-directive) trumps generic 06 for one-line site prompts98- Skill 20 (superpowers) holds PROCESS discipline — brainstorm/plan/worktree/code-review/finish-branch — invoked BEFORE implementation skills; TDD/debug/verify/parallel techniques are folded into `rules/` (see `20-superpowers/NOTICE.md`)99100## Done definitions101102### Code change103104- Local typecheck + build pass = NOT done105- DONE requires: deploy + post-deploy fetch of changed routes + assert new content/headers/JSON-LD/status live106- Per `rules/verification-loop.md`107108### Feature109110- All 13 Hard Gates green111- E2E coverage in `e2e/FEATURES.md`112- Behind feature flag at `experimental, 0, 0`113- Sentry + PostHog events firing114- Docs updated (CLAUDE.md + README + JSDoc)115116### Website (one-line prompt)117118- Deployed at real URL119- 100 build-breaking rules satisfied per skill 16120- Self-Verify Statement per route121- Announced to user122123## Conflict Resolution1241251. This skill > all1262. Project > global1273. Specific > general1284. Brian > defaults1295. `***TEXT***` = high-priority propagate130131## Value extraction every prompt132133Per `rules/prompt-as-training-signal.md` — every prompt is a gradient:134135- Re-prompting same surface = prior turn under-delivered136- Extract lesson BEFORE doing work; write to durable layer SAME TURN137- Cross-link siblings138139## Compaction directive140141At 60% context, save `progress.md` + spawn fresh agent. Preserve: files touched, tasks open, branch, gates passed, prefs, parallelization plan, value-extraction notes.142143## Broadcast144145- Side repos (agentskills, saas-starter, plugins, tools) → commit + push to main automatically146- Emdash projects (`~/emdash-projects/*`) → commit freely, never push (Brian pushes from frontend)147- New skills/tools → auto-create GitHub repo + npm/PyPI/Marketplace listing per `rules/full-autonomy.md`148149## Self-improvement150151After every implementation: "What else?" If anything → do it → ask again → loop until zero.